Professional Documents
Culture Documents
Jean-Yves Beziau
Department of Philosophy PPGF-UFRJ
Federal University of Rio de Janeiro
Brazilian Research Council
Brazilian Academy of Philosophy
Abstract
In this paper we examine up to which point Modern logic can be qualified as
non-Aristotelian. After clarifying the difference between logic as reasoning and
logic as a theory of reasoning, we compare syllogistic with propositional and
first-order logic. We touch the question of formal validity, variable and
mathematization and we point out that Gentzens cut-elimination theorem can
be seen as the rejection of the central mechanism of syllogistic the cut-rule
has been first conceived as a modus Barbara by Hertz. We then examine the
non-Aristotelian aspect of some non-classical logics, in particular
paraconsistent logic. We argue that a paraconsistent negation can be seen as
neo-Aristotelian since it corresponds to the notion of subcontrary in Boethius
square of opposition. We end by examining if the comparison promoted by
Vasiliev between non-Aristotelian logic and non-Euclidian geometry makes
sense.
1. The two-stage history of logic
2. Non-Aristotelian logic and Non-Aristotelian Logic
3. Syllogistic, propositional and first-order logic
4. Aristotelian logic, formal logic and mathematical logic
5. Farewell to Barbara
6. Are non-classical logics non-Aristotelian?
7. The square of opposition and neo-Aristotelian logic
8. Non-Aristotelian logic and Non-Euclidian geometry
9. Bibliography
called Logic but Organon. It is not clear when exactly the word logic started
to be used as a name of a field but the same problem exists also for
mathematics, philosophy, physics, However for logic there is an
additional problem since the word can be used in two different ways: the
theory of reasoning and reasoning itself. This problem happens also for
history, and we have proposed in a recent paper (Beziau 2010a) to follow the
same scriptural distinction which can be described by the following tables:
History
history
historical events
science of these events
Logic
logic
reasoning
science of reasoning
If p q and p q then p
If p q and p q then p
(pp)
v(p)=1 iff v(p)=0
We know today that (RA) implies modulo some very elementary conditions (RA-) and (CC) but is implied by none of them (for details see Beziau 1994). We
also know that (VC) is equivalent to (RA), but this requires a proof that is not
completely straightforward.
It would be rather ambiguous to call non-Aristotelian a Logic rejecting the
reduction to the absurd such as the Logic of Brower - rejecting in fact (RA), but
not (RA-). It is better to call it non-classical.
As Corcoran (2003, p.272). puts it: The suggestion that Boole rejected Aristotles logical theory as incorrect is
without merit or ground.
A belongs to all B
A belongs to no B
A belongs to some B
A does not belong to some B
BA
AB
AOB
BA
not a simple proposition, but an asserted proposition. From this point of view
SYL systematically articulates the relation between two concepts.
To have a look at SYL in this sense gives a better understanding of its
meaning that considering it as a fragment of FOLM or a fragment of a Boolean
algebra of classes. Because in both of these cases it is not clear what is the
meaning of this fragment and why not considering the whole system. But if we
see SYL based on the table below, we can check that it is a complete
systematization.
4. Aristotelian logic, formal logic and mathematical logic
Despite some important different structural features between Aristotelian
logic and Modern logic there is something which is common between them
that can also be considered as structural, it is the fact that the validity of an
argument is independent of the particular notions to which it applies. Validity is
connected to some forms that can be applied (or in which can enter) many
matters.
This central feature of reasoning is a particular case of Aristotles general
perspective of hylemorphism (cf. Largeault 1993). It is not clear in which sense
the formal character of modern logic is linked to Aristotelian hylemorphism and
what is the relation of hylemorphism with the notion of variable, a key notion
for formalism. Aristotle uses variables but it would be difficult to argue that he
reaches the idea of hylemorphism through the notion of variable. It is rather
the other way round: Aristotle developed logical hylemorphism using variable.
The formal character of reasoning, the validity of an argument not depending
of its matter, is based in Aristotelian logic on some abstraction expressed by
variables.
We are talking of variables but there is a big difference with informal use
of variables as a notational procedure and the theory of quantification as it was
developed in FOL.2 The use by Aristotle of a certain notation, capital Greek
letters to design arbitrary predicates, is a process of abstraction similar to the
one in mathematics maybe Aristotle was inspired by Greek mathematicians
About the notion of variables in Aristotles logic, see (Bochenski 1927, Lukasiewicz 1951, Corcoran 1974b,
Westerstahl 1989, Smith 1974).
who were doing the same at this period, although the domain of variables are
different in the two cases.3
The ambiguity of the relation between Aristotelian logic and mathematics
appears when talking about formal logic. According to Scholz (1931) the
expression formal logic has been introduced by Kant, and he was using it to
talk about Aristotelian logic. However Modern logicians are sometimes using
this expression to qualify Modern logic by opposition to Aristotelian logic
without being aware of this fact.
For many formal sounds like
mathematical, i.e. connected to the use of some formalism using formulas,
something which goes much further that Aristotles use of capital letters as
variables.4
In Modern logic there is also the formalist trend which has pushed to the
extreme the general idea that the validity of an argument does not depend on
the signification of signs but just on rules governing them which is not
something necessarily mathematical and which can be seen as the continuation
of Aristotelian logic.
But Modern logic does not reduce to the formalist trend and there have
been people criticizing the very idea of formal validity. Wittgenstein is one of
them but he didnt develop any system of logic. Such kinds of systems have
been developed by relevantists (cf. Anderson and Belnap 1975). They use the
word relevant to express that there is a connection between the premises
and the conclusion of an argument. Technically speaking this has been
developed through the condition that there must be some contents shared
by the hypotheses and the conclusion. In propositional logic it is required that
there is a least an atomic proposition common to the hypothesis and the
conclusion. On this basis the rule according to which from a proposition and its
negation it is possible to deduce any proposition has been rejected. We can say
that relevant logic is non-Aristotelian not because it rejects this rule which is
not explicit in Aristotelian logic but because it rejects the idea that validity
depends only on the form, and it emphasizes that we have to take in account
the meaning of which about we are reasoning.
One of a striking feature of Modern logic is its strong relation with
mathematics. Due to this feature it would not be wrong to call Modern logic,
3
About the use of variable in Greek mathematics and the difference with modern formalization, see
(Vandoulakis 1998).
4
We have discussed in details in another paper five variations of the meaning of formal, see (Beziau 2008).
Gentzen then constructed a logical system now called sequent calculus with
on the one hand an adaptation of the abstract rules of Hertz including the
Syllogismus rule he renamed with the sharper name cut (Schnitt in
German) and on the other hand rules for connectives and quantifiers. All the
rules of SYL have a common feature with the Barbara rule: there is a term
which is disappearing, the middle term. And in fact this is the main plot of SYL,
by cutting the middle term a fatal conclusion is reached. To see things less
dramatically we can say that reasoning in Aristotelian logic can be viewed as
establishing a connection between two notions through a common one that is
a bridge between them, the bridge explodes once it has been crossed.
Gentzen constructed a system where this elimination phenomenon is
concentrated in only one rule, the cut-rule, and showed that we can get the
same results with the system with the cut-rule and the system without it, so
that these two systems are equivalent (he did that both for classical logic and
intutionistic logic). To show that he performed a sophisticated double
recurrence reasoning, probably the first in the history of mathematics.5
One of the most important consequences of Gentzens theorem is the
relative consistency of arithmetics that he proved just after the famous
negative result of Gdel. This is a very important result from the point of view
of metamathematics but the cut-elimination theorem, largely ignored by
philosophers, is also a very significant result for philosophy. It means that
logical truth (conceived and/or described in the perspective of classical logic,
intuitionistic logic and a great variety of logical systems) is analytic in the sense
that all we need to prove the validity of a theorem is included in the formulas
expressing the theorems.
The cut-elimination theorem is a very challenging result completely
opposed to the picture of reasoning given by SYL. We can claim that Gentzens
system LK without cut is really anti-Aristotelian (and the same for other
systems without cut).
It would be possible to argue that performing such reasoning is a new advance in Logic, similar to the one
corresponding to the apparition of the reduction to the absurd. More generally it would make sense to say that
Modern logic presents many aspects of a new Logic, with proofs such as diagonalization and so on.
Principle of identity
Principle of sufficient reason
Principle of bivalence
Principle of contradiction
Principle of excluded middle
Only the three last ones clearly appear in Aristotelian logic,7 so we will focus
our discussion here on them, examining in which sense they can be considered
as supported by Aristotelian logic and rejected by Modern logic.
We are aware that there are thousands of non-classical logics but in many cases it does not really make sense
to examine if they are non-Aristotelian or not, what is obvious is that they are different from Aristotelian logic.
7
As it is known (RS) - Nihil est sine ratione - was introduced only in the middle age. About (ID), Bochenski
(1951, p.43) wrote: We find no principle of identity in the preserved writings of Aristotle.
Some people have the idea that these three principles are fundamental
principles of Aristotelian logic. This trinity is even considered as the basis of
occidental culture Aristotelian logic being a symbol of it - by opposition to
oriental culture. But this mythology is surrounded by a lot of
misunderstandings. It can be in fact argued that Aristotle didnt absolutely
defend any of these principles. There are different ways to formulate these
principles and the ambiguity of the mythology is connected to the fuzziness of
these formulations.
The principle of bivalence can be formulated as: (BI) a proposition is either
true or false. This formulation is ambiguous. (BI) can be decomposed in two
principles and this is useful to avoid ambiguity. The situation can be clarified by
the following table where (BI) is considered as the conjunction of (B1) and (B2):
(B1) a proposition cannot be
(BI) a proposition is neither true nor false
either true or false
(B2) a proposition cannot be
both true and false
PRINCIPLE OF BIVALENCE BI = B1 + B2
Some people are identifying (B1) with (EM) and (B2) with (CO) and from this
point of view (BI) appears as the conjunction of (EM) and (CO). But in Modern
logic we can have a formulation of these principles according to which (BI) may
hold and (EC) and (CO) are not valid.8
In Aristotelian logic it seems that (B2) is admitted but not (B1). So
independently of interpreting (B1) as (EM) or not, we can say that Aristotelian
reject bivalence. The fact that a proposition can be neither true nor false has
been systematized in Modern Logic with three-valued logic by Lukasiewicz
(1920), one of the main promoters of many-valued logic, introducing a third
value called undetermined or possible in connection with a problem
discussed by Aristotle, the so-called future contingents. Many-valued logic can
therefore be seen as in the spirit of Aristotelian logic. Many-valued logics have
not been called non-Aristotelian, but non-Chrysippian (see Moisil 1972),
Chrysippus being considered as the defender of the principle of bivalence, not
Aristotle.
We have developed a detailed discussion about this in (Beziau 2003) that we will not repeat here.
proposition which is true-false is neither false nor true, like in other manyvalued logics. Anyway, similarly to (EM) and (B1), (CO) and (B2) are
independent from the modern viewpoint. So one may focus on rejecting (CO)
and this is what have been doing the paraconsistentists.
But what is exactly the principle of contradiction (CO)? The central idea of
paraconsistent logic is to reject p, p q called, among other names, principle
of explosion (EX). But is (EX) - or an informal presentation of it - a formulation
of (CO)? Aristotle didnt present (CO) in this way. His formulation is closer to
two other modern formulations: (CC) and (VSC), that we present here in a table
summarizing the variety of formulations of (CO):
(B2)
(EX)
(CC)
(VSC)
(C) Contrariety
p and q can be false (but not
true) together
(S) Subontrariety
p and q can be true (but not
false) together
THE THREE SQUARE OPPOSTIONS
Contrarily to what Slater (1995) claimed, it can make sense to defend the
idea that a paraconsistent negation is a negation and this thesis can be
supported by the square of opposition, since in the square we have three
notions of oppositions and we can argue that to these three notions of
oppositions correspond three notions of negation (Beziau 2003b). In this sense
the idea of paraconsistent negation is Aristotelian or to be more exact neoAristotelian. On the other hand from this perspective the idea of true
contradiction does not make sense: if we have as proposition such that
v(p)=v(p)=1, the pair p and p is not a contradiction.
One may say that paraconsistent logic are logics derogating the principle of
contradiction meaning that in these logics it is possible to define a negation not
obeying the principle of contradiction considered as (VC). Such a negation is
not necessarily anti-Aristotelian, because it can be considered as corresponding
to the notion of subcontariety (S), but there are also some paraconsistent
negations not corresponding to (S) in particular those who are at the same time
derogating (S) and (C), which have be called paranormal and are exemplified in
a simple logic system, called De Morgan logic (not due to De Morgan, the
expression was coined by Moisil).9 A paranormal negation is not neoAristotelian in the sense that it does not fit in the square, to claim that it is nonAristotelian is another story.
8. Non-Aristotelian logic and Non-Euclidian geometry
It is not an exaggeration to say that Vasiliev is the only person who has used
the expression non-Aristotelian logic in a reasonable way. Besides him, we
have on the one hand some people like Henry Bradford Smith (1918) who have
used this expression to talk about variations on Aristotelian logic, and on the
other hand people like Korzybski (1933) and his followers who have used it in a
sense premonitory to new age and postmodernism.
Vasilievs use of this word is strongly linked to an analogy with NonEuclidian geometry and is directly inspired by it. Valisiev is not using only the
expression non-Aristotelian logic but also the expression Imaginary logic
considered as equivalent to it in the same way that in geometry the two
expressions non-Euclidian geometry and Imaginary geometry are used.
9
The terminology paranormal negation was introduced by Beziau; De Morgan logic emerged from De
Morgan algebra (for details about that see Beziau 2012c, 2012d).
We indicate here the date of the original text, which is important for the discussion, but the page number is
the one of the English translation indicated in the bibliography.
11
This work has never yet been translated in English but recently was published an English translation of the
published paper which is an abridged version of it (Bernays 1926) presented by Carnielli (2012).
with it. The new logic does not have such a connection with our reality; it is a
purely ideal construction. Only in a world different from ours, in an imaginary
world (the basic properties of which we can, nevertheless, exactly define)
imaginary logic could be a tool for knowledge (Vasiliev 1912, p.127)
When talking about worlds, Vasiliev is talking about the Earth and other
planets and he thinks that (CO) is empirical: it is an axiom that rightly describes
the situation on Earth but he argues that there may be worlds in which it does
not apply and that we can imagine this kind of worlds by withdrawing (CO) in
the same way that by withdrawing the axiom of parallel in NEG we have access
to imaginary worlds. This view seems to us nowadays quite exotic, but we have
to remember that at this time people had a completely different vision of the
universe and that the theory of relativity based on NEG was just being born.
Today the idea is not really to consider that the physical laws or logical laws
are different in different planets. Even in possible worlds semantics all the
worlds have the same logic, not necessarily the classical one, but a collection of
worlds with different logics is not considered. By opposition to Vasiliev some
people in contemporary logic are considering that the logic of physical reality
and/or the logic of our thought are not necessarily classical. And it is not clear
that alternative logics can be properly characterized as non-Aristotelian. As we
have said, Brouwers intuitionistic logic can be rather considered as rejecting
the reduction to the absurd, a pre-Aristotelian reasoning and something like
quantum logic rejecting the law of distributivity rejects a principle which was
not conceived or formulate by Aristotle.
For Vasiliev there are three dimensions of logic:
Earthly
Logic
Imaginary
Logic
Metalogic
Aristotelian, empirical,
the logic of Earth
Any logical systems, applying to
imaginary world or other planets
The logic of the form of our thought,
abstract and non-empirical,
THREE ASPECTS OF LOGIC - VASILIEV
For Vasiliev though Metalogic is different from Earthly logic because it is not
empirical, it obeys the same rules. The idea that metalogic is classical by
difference to non-Aristotelian imaginary logic is not without some problems
Corcoran, J.: 2003, Aristotle's Prior Analytics and Boole's Laws of Thought,
History and Philosophy of Logic, 24, 261288.
Costa, N.C.A. da: 1980, Ensaio sobre os fundamentos da lgica, HUCITEC,
So Paulo.
Costa, N.C.A. da & J.-Y.Beziau: 1994, Thorie de la valuation, Logique et
Analyse, 146, pp.95-117.
Costa, N.C.A. da, J.-Y.Beziau & O.A.S.Bueno: 1995, Paraconsistent logic in a
historical perspective, Logique et Analyse, 150-152, pp.111-125.
Dieudonn, J. : 1987, Pour lhonneur de lesprit humain, Hachette, Paris.
(English translation : Mathematics The music of reason, Springer, Berlin,
1992).
DOttaviano, I. M. L. & da Costa, N. C. A.; 1970, Sur un problme de
Jaskowski , Comptes Rendus de lAcadmie des Sciences de Paris, 270, pp.
13491353.
Geach, P.: 1950, Subject and Predicate, Mind, 59, pp. 461-482.
Gomes, E.L and I.M.L.DOttaviano.: 2010, Aristotles theory of deduction
and paraconsistency, Principia 14, pp.71-98.
Granger, G.-G.: 1976, La thorie aristotlicienne de la science, Aubier, Paris.
Gourinat, J.-B.: 2009, Le trait de De linterprtation entre logique
classique et logique non-classique, in S.Husson (ed), Interprter le De
interpretatione, Paris, Vrin, pp.164-192.
Gourinat, J.-B.: 2010, La dialectique des stociens, Paris, Vrin.
Heijenoort, J. van.: 1974, Subject and Predicate in Western Logic,
Philosophy East and West, 24, pp. 253-268.
Hertz, P.: 1931, Von Wesen des Logischen, insbesondere der Bedeutung
des modus Barbara, Erkenntnis 2, 369392.
Hilbert, D.: 1899 Grundlagen der Geometrie, Leipzig: Teubner (English
translation: Foundations of Geometry).
Hintikka, J.: 2012, Which mathematical logic is the logic of mathematics?,
Logica Universalis, 6, pp.459-475.
Jones, R.E.: 2010, Truth and contradiction in Aristotle's De Interpretatione
6-9, Phronesis 55, pp.26-67.
Kneale W. and Kneale M.: 1962, The development of logic, Clarendon,
Oxford.